Ever heard of "regular expressions"? Robo* is a much easier word to type than Robolito. And "most" here understand exactly what I mean. Very similar to the texters typing "where R U?" Most understand the "R U" without needing an interpreter.Rolf wrote:From the perspective of beancounter facts. But not the perspective of the whole.bob wrote: This is not a "beancounter myth" as you would like it to be. It is an absolute statement of fact.That's a typical statement of beancounter quality. Someone said that something is a clone, that is correct. But was this about everything else? Was it about Robo* at all? And for all what do you intend with the funny abrevation Robo*? Do you want to maintain the impression of a minimum of non-hatred research caution on your side in relation to Vas?As far as the Robo* issue, all that has to happen is for Vas to provide evidence that the program is reverse-engineered from Rybka. And all of the Robo* talk will go away. But the silence is deafening, and by the same token of fairness, one can not ostracise Robo* with no evidence to support that action. And we have absolutely nothing other than a single statement by the author of Rybka that it is a clone, with no evidence of any kind to support that.
There's no hatred in my following this topic, only an interest in knowing what really happened. I now have a pretty good idea.
You do realize that if I release my source code, it is not "public domain"??? It is copyrighted and the copyright is held by me unless I explicitly forgo that copyright, as is done when one chooses to release under the GPL. So copying fruit code would _never_ be acceptable. Never. not ever. Think you can grasp that simple point one day???Professor! Sorry. Excuse me, the quote, the true quote, so much time must be, is that it contains only original code plus public domain. Please research your data correctly before you make such offending statements. Excuse me Bob, but you are now under examinig eyes yourself, so please try to be correct."Decent Gentleman" said there was absolutely no fruit code in Rybka 1. That is plain and simple false.
I'd be willing to bet there is no communication between you and Vas except for stuff you read that he wrote, and then you interpret what he wrote in some of the most bizarre translations known to mankind.Ok, there you go again. I thought you were a famous beancounter who has his facts straight. The statement is premature because you forget that we are talking about business matters. And yes, I admit Vas should have notified to the University of Alabama why he had marked the clones with the label clone. Certainly Vas will be better prepared in the next clone affair, because that was really unnecessary negligence. You at least had deserved to get the full proof against the clones. Vas will change that after I told him that you were the patron here. Alone because only this way Vas could avoid the next hate campaign against himself.Now he says that robo* and IP* are clones of Rybka 3, with absolutely no evidence. Of course we should believe him. I gave him a long "benefit of the doubt" as I figured that he would do the normal thing and provide evidence for such a strong claim. -nothing-.
Two issues. (1) he copied code from fruit, which should not have happened. Whether this copied code is still in Rybka today is unknown at present, and this is irrelevant. It was wrong when it was done, plain and simple. (2) he claimed a new program was a clone, and after having been given plenty of time to offer something to support this, he has provided nada. zilch. zippo. nothing. And that is _also_ wrong. If you are going to accuse, you need to provide evidence. It'd be a great world if someone could just walk into the nearest police station and say "xyz did this, arrest him" with no evidence of any kind, and the cops go round him up on the spot. Doesn't work like that with the police, and it shouldn't work like that in computer chess either. We've already had too many clone claims, not all of which were valid. So far, this one appears to be in that latter category as well.
What you confound is that in business and justice, other than the laboratory of science, you dont exchange all the data you have because you wont want to hurt your case. The details however important have no inborn reason to be published as such. If you really were a neutral scientist, sworn your oath on a minimum of discretion, I am sure tht you had got the detsils if you had asked. But NOT in your case when you are already in the Strelka case one of the loudest preachers of prejudice and hate. IMO you are just not trustable in questions of RYBKA. I dont want to speculate about the reasons. First I am a computerchess and programming layman, but secondly this is still a pending case. But the science stuff you can just forget it in your case. The same as in the one of DB2/IBM with all friends of yours. You dont even accepted a minimum of criticism that came from a science angle itself, about the treatment of clients and stuff like that. Suddenly you declared that all wasnt about science but a sport where it's important to win by all means. Like in boxing I might add to mark the main idea in such a claim. Because if you once admit by all means then the truth is somewhere over the rainbow but no more here on Earth. All that of course seen from Europe. Sorry.
So far the beancounter mantra. I dont understand the details but I get what you and your "volunteers" did in disassembling code out of other people's engines. Like you do that all day long with other commercial progs like FRITZ, HIARCS, SHREDDER and JUNIOR. In special JONNY comes to mind because that engine was almost as weak as Rybka in the early tournament that was mentioned in an attempt to harm Vasik.I have seen fruit 2.1 source code. And I have seen disassembled Rybka 1 assembly language. I don't have any problem going from assembly language to C and back. I teach this exact thing _every_ semester in my x86 assembly language programming course. Any other questions you have?
(The whole code was examined or just some pieces?) Pieces. If pieces of Fruit are copied, the GPL is violated. You don't have to copy the entire thing. What part of that do you fail to understand?? (Because how do you know that what you saw in Rybka 1 beta sourcecode (??) really can only come out of Fruit?) From 40 years of programming experience, which has clearly shown how unlikely it is for two students to produce the same code for very simple assignments, much less for something as complex as a chess engine. I have explained this previously. Multiple times. My answer is not going to change. All you need to do is take Crafty Source, and (say) Fruit 2.1 source and compare them looking for identical blocks of code, or program structure, or data structures, or anything. And when you find zero, then compare crafty to gnuchess, or another open-source engine. And keep going until you either find identical blocks of code or give up. If you find identical blocks of code, we will immediately contact the author of that program as he probably asked me if he could use parts of Crafty and I agreed. But look first, then you will see the problem. Give a simple assignment to an English class. "Write a two page description of XXX." XXX can be any topic that is current. And then compare the write-ups for duplicate wording anywhere. It just doesn't happen. Unless something is copied. (Just for my negotiations with legal authorities. Would that be possible if you related to a page or a message here, or just a quote you post here now, so that I could have the data for my questionnary?) I'm not going to go back thru past CCC posts to do your research. The topic was discussed _at length_ last year. On several occasions. (Again, I dont doubt your expertise, because that would be ridiculous, but it must be allowed to question the sources that made you so confident that you could compare true code. Bob, and please do not play the Strelka card. It proves nothing. Just the codes you have examined.) I have looked at all 3. I first compared Fruit to strelka and saw similarities that looked suspicious. Some claimed foul, saying that is not really Rybka even though Vas explicitly claimed it as his code, originally. So we had some volunteers that disassembled parts of Rybka 1's binary executable, so that the strelka criticism would no longer be valid. But the same problem was still there. Duplicated code, names, bizarre programming practices, etc. (Worldwide also here in Germany in a famous forum people want to know if it is really possible that you had the original source code of the two mentioned relevant programs before your eyes. There are some who said this is impossible. Others wrote yes, but Bob had only the UCI code. Etc. Therefore I ask you here. I'm just the messenger, please dont think that I made this all up.)
We simply looked at the disassembled executable, which turns into assembly language. It is not a complex task to turn assembly language directly into C, any more than it would not be difficult to turn C into assembly language by hand. I teach students to think this way in my assembly language course, in fact.
But I didnt ask these questions above to become a real expert in computer science but because I am on a sort of interdisciplinary trip to find out potential weaknesses in your approach as such.
Here now my condensated question to reveil first weaknesses:
In your attempt to prove wrongdoing in Rybka, Bob, you took code from Fruit and code from Rybka, right? And you found similaritiess or even identical pieces of code. I want to believe that for the following proof of your failure. So, here comes now. How could you know that Vas took that code from the Fruit 2.1.? How did you examine if it wasnt from public domain? How did you secure that that whyt ou have found, resulted from a wrong taking code out of that GPL code? How do you want to prove that without having spoken to Vas? Do you think that you can make a case based on evidence, against the loud statement of your collegue that he had only original code in Rybka plus public domain? How do you discard OTHER explanations - say other provenience for these parts of code that you have researched, when you admitted that you havent examined the complete source code at all because you simply dont have it in total? What if Vas got a certain copy of Fruit by Fabien in late 2005? What if he bought some code from Fabien? What is if this happened before the code was declared GPL? Have you researched, what would be a trivial duty for a beancounting scientist, the complete history of the months in late 2005? Between Fabien and Vas?? How can you be so damn certain that you think you are entitled to claim that Vas just plain lied if he said that Rybka contained ONLY original code plus public domain? Can you prove that he lied in a legal case before a court in the USA?
Just a couple of questions. Do you answer some of them?
Still in all respect for your 40 years of top expertise in computerchess.
Rolf
